stable sulfur isotope fractionation and discrimination between the sulfur atoms of thiosulfate...

$: Stable sulfur isotope fractionation and discrimination between the sulfur atoms of thiosulfate during oxidation by Halothiobacillus neapolitanus$
R E S E A R C H L E T T E R

Stable sulfur isotope fractionationanddiscrimination between thesulfuratomsofthiosulfate during oxidation byHalothiobacillusneapolitanusDonovan P. Kelly

Department of Biological Sciences, University of Warwick, Coventry , UK

Correspondence: Donovan P. Kelly,

Department of Biological Sciences, University

of Warwick, Coventry CV4 7AL, UK. Tel.: 144

0 24 7657 2907; fax: 144 0 24 7652 3701;

e-mail: [email protected]

Received 31 October 2007; accepted 25

February 2008.

First published online 28 March 2008.

DOI:10.1111/j.1574-6968.2008.01146.x

Editor: Christiane Dahl

Keywords

stable sulfur isotopes; 34S/32S ratios;

Halothiobacillus ; isotope discrimination;

thiosulfate oxidation.

Abstract

Growing cultures and nongrowing suspensions of Halothiobacillus neapolitanus

selectively fractionated 32S and 34S during the oxidation of the sulfane- and

sulfonate-sulfur atoms of thiosulfate. Sulfate was enriched in 32S, with d34S

reaching � 6.3% relative to the precursor sulfonate-sulfur of thiosulfate, which

was progressively resynthesized from the thiosulfate-sulfane-sulfur during thio-

sulfate metabolism. Polythionates, principally trithionate, accumulated during

thiosulfate oxidation and showed progressive increase in the relative 34S content of

their sulfonate groups, with d34S values up to 120%, relative to the substrate

sulfur. The origins of the sulfur in the sulfate and polythionate products of

oxidation were tracked by the use thiosulfate labelled with 35S in each of its sulfur

atoms, enabling determination of the flow of the sulfur atoms into the oxidation

products. The results confirm that highly significant fractionation of stable sulfur

isotopes can be catalyzed by thiobacilli oxidizing thiosulfate, but that differences in

the 34S/32S ratios of the nonequivalent constituent sulfur atoms of the thiosulfate

used as substrate mean that the oxidative fate of each atom needs separate

determination. The data are very significant to the understanding of bacterial

sulfur-compound oxidation and highly relevant to the origins of biogenic sulfate

minerals.

Introduction

Naturally-occurring sulfur contains four stable isotopes, the

most abundant being 32S (95%) and 34S (4.2%), with the

ratio of the two in biological materials varying significantly

as a result of enzymatic discrimination, often in favour of

the lighter isotope. The standard procedure to measure

isotope discrimination is to measure the 34S/32S ratio and

to express deviation from an international standard (IAEA-

S-1; Krouse & Coplen, 1997) as d34S (in parts per thousand;

%), using the following equation:

d34S ¼ 34S=32S� �

sample= 34S=32S� �

standard

h i� 1� 103

Extensive studies have been made of stable sulfur isotope

fractionation during sulfide production by sulfate-reducing

bacteria, largely because of the geomicrobiological signifi-

cance of the biogenic formation of stratified sulfide minerals

(Kaplan & Rafter, 1958; Jensen, 1965; Chambers et al., 1975;

Fry et al., 1988; Bottcher et al., 1999; Detmers et al., 2001;

Knoller et al., 2006). Sulfate-reducing bacteria produce

sulfide in which discrimination in favor of 32S has been

observed to produce d34S values of � 7% and � 42%, and

may even exceed � 70% (Chambers & Trudinger, 1979; Fry

et al., 1988; Smock et al., 1998; Hoek & Canfield, 2008).

Much less is known about isotope discrimination during

reduced sulfur compound oxidation by chemolithotrophic

bacteria (Kaplan & Rafter, 1958; Kaplan & Rittenberg, 1964;

Chambers & Trudinger, 1979; Karavaiko et al., 1981; Fry

et al., 1986).

In the most recent study, over 20 years ago, Fry et al.

(1986) reported negligible sulfur isotope effects during

thiosulfate oxidation by Thiobacillus versutus (now renamed

Paracoccus versutus). This contrasted with some earlier work

on Thiobacillus concretivorus (now renamed Acidithiobacil-

lus thiooxidans) in which sulfide oxidation resulted in 32S

enrichment in sulfate (d34S � 10.5% to � 18.0%), and the

accumulation of polythionate (SnO62�) enriched in 34S (d34S

FEMS Microbiol Lett 282 (2008) 299–306 c� 2008 Federation of European Microbiological SocietiesPublished by Blackwell Publishing Ltd. All rights reserved

up to 119%; Kaplan & Rittenberg, 1964). Since 1986, major

reclassification of the Thiobacillus genus has occurred and

variant pathways of sulfur-compound oxidation have been

recognized among its former species. A unitary mechanism

for thiosulfate oxidation has been established in the faculta-

tively chemolithotrophic Alphaproteobacteria such as Para-

coccus, Starkeya and Pseudaminobacter (Lu et al., 1985; Kelly,

1989; Kelly et al., 1997; Friedrich et al., 2001; Kappler et al.,

2001; Quentmeier et al., 2003; Friedrich et al., 2008; Sauve

et al., 2007), but the inorganic sulfur compound oxidation

mechanisms operating in most chemolithotrophic Betapro-

teobacteria and Gammaproteobacteria are not yet fully

elucidated (Kelly, 1989, 1999; Kelly & Wood, 1994a, b; Beller

et al., 2006). The thiosulfate-oxidizing multienzyme Sox

system operates in the Alphaproteobacteria (Bamford et al.,

2002; Quentmeier et al., 2007; Reijerse et al., 2007; Sauve

et al., 2007), but bacteria such as Thiobacillus and Halothio-

bacillus contain at most only some genes encoding the Sox

proteins; they also commonly produce polythionates during

thiosulfate oxidation, and grow readily on tetrathionate and

trithionate (Wood & Kelly, 1986; Kelly, 1999; Petri et al.,

2001; Kelly & Wood, 2005; Kelly et al., 2005; Beller et al.,

2006; Meyer et al., 2007).

A role for isotope discrimination during sulfur com-

pound oxidation was strongly indicated by the finding that

the jarosite and gypsum (iron and calcium sulfates) sur-

rounding altered pyrite deposits showed the sulfates to have

d34S values of � 7% to � 10% relative to the pyritic-sulfur,

consistent with their formation by bacterial oxidation of the

sulfide to sulfate (Nissenbaum & Rafter, 1967).

It was thus considered essential to reassess isotope

discrimination during thiosulfate oxidation by the obligate

chemolithotroph Halothiobacillus neapolitanus, which was

previously reported to produce sulfate and polythionate

enriched, respectively, in 32S and 34S (Chambers & Trudin-

ger, 1979). Detection and interpretation of any stable sulfur

isotope effects was expected to be complicated by the known

resynthesis of thiosulfate and of trithionate from the

sulfane-sulfur atom (� S�) of thiosulfate (� S� SO3�)

during its oxidation (Trudinger, 1964a; Kelly & Syrett,

1966; Kelly & Wood, 1994a). Interpretation is further

complicated by the fact that the two sulfur atoms of

thiosulfate are neither chemically equivalent (S-oxidation

states of � 1 for the sulfane-atom and 15 for the sulfonate-

sulfur; Vairavamurthy et al., 1993) nor isotopically equiva-

lent with respect to 34S/32S ratios. Different batches of

commercially available sodium thiosulfate show d34S values

ranging from about 14% to � 4% (sulfane-sulfur) and

14% to 115% (sulfonate-sulfur), typically with a differ-

ence of 6–14% between the two atoms (Chambers &

Trudinger, 1979; Fry et al., 1986; Smock et al., 1998).

Chambers & Trudinger (1979) suggested that a fruitful

approach to resolving the problem of discrimination in

34S/32S fractionation during the oxidation of each sulfur-

atom of thiosulfate might be to use thiosulfate labelled in

one or other atom with 35S, enabling correlation of stable

isotope patterns with the transformations of the individual

sulfur atoms. This would also enable tracking of the role of

polythionates. The differential use of the stable sulfur

isotopes during thiosulfate oxidation by three strains of

H. neapolitanus has been assessed, and 35S tracers used to

assist in discriminating between the effects seen with the

nonidentical sulfur atoms within the thiosulfate.

Materials and methods

Organisms and growth conditions

Halothiobacillus neapolitanus strain C (DSM 581), strain X

(ATCC 23641) and strain FA11 (Kelly, 1968) were main-

tained on thiosulfate agar mineral medium and grown in

liquid medium with 40 mM sodium thiosulfate as described

previously (Kelly & Wood, 1998). Cultures (c. 1 L) for

isotope fractionation experiments were grown at 30 1C with

forced aeration with air and automatic pH control at pH 6.5

with sodium bicarbonate. For experiments with nongrow-

ing suspensions, cultures were harvested by centrifugation,

washed with 0.9% NaCl and resuspended in 0.1 M potas-

sium phosphate buffer, pH 7.0.

Sampling of growing cultures for stable isotopefractionation, and recovery, determination anddegradation of sulfur compounds

Cultures were inoculated with actively growing organisms

and were monitored for thiosulfate consumption for up to

13.5 h. Samples were removed at intevals, organisms re-

moved by centrifugation, and the supernatant solutions

assayed for their sulfate, thiosulfate and polythionate

(SnO62�) content before further treatment and degradation

for determination of 34S/32S ratios of sulfur compounds.

Sulfate was estimated by barium precipitation, and thiosul-

fate and polythionates by spectrophotometric analysis after

cyanolysis (Kelly & Wood, 1994b).

Sulfate in the supernates was recovered after precipitation

with barium chloride, normally in acid solution with HCl

and isolated by centrifugation or filtration. Thiosulfate used

as the substrate and polythionates and residual thiosulfate in

supernates, were degraded by reaction with mercuric chlor-

ide as described by Kelly & Wood (1994b). This procedure

precipitated the sulfane-sulfur [� S� ] of thiosulfate and

poythionates as a mercury complex, and left the sulfonate-

sulfur [� SO3�] in solution as sulfate (Van der Heijde &

Aten, 1952). The latter was precipitated with barium chlor-

ide at room temperature in the presence of acetic acid. The

mercury complex of the sulfane-sulfur was oxidized to

sulfate with bromine-saturated concentrated nitric acid in

FEMS Microbiol Lett 282 (2008) 299–306c� 2008 Federation of European Microbiological SocietiesPublished by Blackwell Publishing Ltd. All rights reserved

300 D.P. Kelly

the presence of NaCl, evaporated to dryness on a sand bath

and the residue dissolved in distilled water. The sulfate

produced was recovered after boiling precipitation with

BaCl2 under acid conditions.

In one experiment, residual thiosulfate, sulfate, trithio-

nate and tetrathionate were isolated by ion exchange chro-

matography using Dowex 1� 2 resin, and eluting sulfate

and thiosulfate with 1 and 2 M ammonium acetate, pH 5.0,

respectively, and trithionate and tetrathionate with 3 and

6 M hydrochloric acid, respectively.

To follow the time course of the distribution of the

sources of the sulfur atoms in different sulfur compounds,

thiosulfate labelled with 35S in either the sulfane or sulfonate

atom were used: [35S� 32SO3]2� and [32S� 35SO3]2�, using

previously described methods (Kelly & Syrett, 1966; Kelly &

Wood, 1994a, b). 35S-labelled sodium thiosulfates were

obtained from Amersham (UK).

Stable isotope fraction by nongrowingsuspensions of H. neapolitanus strain C

Duplicate suspensions (0.7 mg dry wt mL�1; 50 mL) in

250-mL Erlenmeyer flasks were shaken at 30 1C and supple-

mented with 0.02 M sodium thiosulfate. Three parallel

incubations were run simultaneously to monitor thiosulfate

metabolism; these contained proportional amounts of thio-

sulfate and bacteria and were: (i) duplicate Warburg flasks

for real time measurement of thiosulfate oxidation; shake

flasks with (ii) [35S� 32SO3]2� to monitor sulfane-sulfur

conversion to sulfate and (iii) [32S� 35SO3]2� to monitor

sulfonate-sulfur conversion to sulfate. After 25 min, activity

in the flasks for 34S/32S analyses and the 35S flasks were all

terminated by addition of equivalent volumes of ethanol,

and the bacteria removed by centrifugation. Supernates were

analyzed as described above.

Conditions tested to ensure that stable sulfurisotope fractionation did not occur during theanalytical procedures

A sample of aqueous sodium thiosulfate was oxidized to

sulfate with bromine, brought to a standard concentration

of 0.05 M sulfate, then supplemented with 45 mM potas-

sium phosphate, pH 7.0 and 42% (v/v) ethanol. Sulfate

was recovered from this mixture in eight different ways:

(1) boiling with just sufficient BaCl2 under acid chloride

conditions to precipitate all the sulfate, then centrifuged;

(2) as in (1) but BaSO4 recovered by filtration through

47 mm Millipore filters; (3) precipitation with BaCl2 at room

temperature in the presence of acid chloride and picric

acid, boiled for 10 min, left 1 h at 20 1C, then filtered;

(4) precipitation at room temperature with BaCl2 and acetic

acid, 1 h, then filtered; (5) as (1) but omitting phosphate and

ethanol from the test solutions; (6) as (1) but only sufficient

barium added to precipitate one-third of the sulfate; (7) as

(1) but sufficient barium added to precipitate two-thirds of

the sulfate; and (8) as (1) but with a one-third excess of

BaCl2. Results are given in the text.

To test the mercury degradation method, 36 mM thiosul-

fate was degraded for 1.5 h at 32 1C or 1 h at 100 1C, and34S/32S determined for the separated sulfane- and sulfonate-

sulfur.

Determination of stable sulfur isotope ratios

Sulfate obtained from bacterial oxidation of thiosulfate and

from the mercury degradation procedure was processed and

analyzed by isotope ratio mass spectrometry by the Institute

of Nuclear Sciences, Lower Hutt, NZ (Friedman et al., 1995).

Additional calculations for d34S are described in the text.

Results

34S/32S composition of the thiosulfate used as asubstrate by H. neapolitanus and test of theisotopic validity of degradation methods

Thiosulfate was completely oxidized to sulfate (50 mM) with

bromine and the sulfate recovered by precipitation with

barium chloride under eight different conditions, including

partial precipitation of only one-third and two-thirds of the

total sulfate (see Materials and methods). No significant

difference in d34S was seen with any method, and the mean

d34S for the combined sulfur atoms of the thiosulfate was

17.0� 0.5% (12 values), relative to the IAEA-S-1 standard.

Degrading thiosulfate by the mercury method at 32 or

100 1C did not affect d34S of the separated sulfane- and

sulfonate-sulfur atoms, which were 14.4� 0.3% (3 values)

and 110.5� 0.5% (3 values), respectively. The 34S/32S ratios

of the individual sulfane- and sulfonate-sulfur atoms of

thiosulfate were used in computations of relative d34S

values, when the oxidation products were derived from the

sulfonate atom only or in varying ratios from the two sulfur

atoms. For clarity, in the following text, d34S values relative

to the IAEA-S-1 standard are expressed as d34S, and those

recalculated relative to the sulfane- and sulfonate-sulfur

atoms of thiosulfate as d34ST.

Assessment of differential oxidation of the twosulfur atoms of thiosulfate by H. neapolitanusdeduced using 35S-sulfane- and 35S-sulfonate-labelled thiosulfate

Time-course experiments showed that strains X and C of

H. neapolitanus both initially formed sulfate about twice as

rapidly from the sulfonate group than from the sulfane-

sulfur of thiosulfate. During oxidation of 35S-sulfane-

thiosulfate, doubly labelled thiosulfate [35S� 35SO3]2� and


301Sulfur isotope fractionation by Halothiobacillus

trithionate [O335S� 35S� 35SO3]2� were formed, and the

rates of formation of these labelled compounds, and of35S-sulfate, and the distribution of the 35S-label within the

labelled thiosulfate and trithionate by suspensions and

growing cultures of H. neapolitanus were determined. From

a number of experiments, 35S initially in the sulfane-atom

disappeared from thiosulfate at 60–80% of the rate of

disappearance of 35S from the sulfonate-labelled thiosulfate,

resulting in complete disappearance of [32S� 35SO3]2� in

about 70% of the time required for the disappearance of

[35S� 32SO3]2�, so that doubly-labelled thiosulfate derived

exclusively from [35S� 32SO3]2�was the only species present

during the period during which the final 30% of the

chemically measurable thiosulfate was consumed. The sul-

fate first formed by suspensions and growing cultures was

initially derived from only from [32S� 35SO3]2�, and at

times when 10% and 15% of the 35S from [32S� 35SO3]2�

had been released as sulfate, only 2% and 6% of the

[35S� 32SO3]2� label appeared as sulfate. Trithionate accu-

mulated during thiosulfate oxidation, with 18–30% of the

initial thiosulfate-sulfur being recovered as trithionate when

all the thiosulfate had all been consumed. 35S-labelling

showed that 75–85% of the trithionate-sulfur was derived

from the sulfane-sulfur of the initial thiosulfate and 15–25%

from the original sulfonate group. These data were applied

to deduce the approximate relative sources of the sulfur in

sulfate and trithionate (and of the combined sulfane- or

sulfonate-sulfur from thiosulfate1polythionate or polythio-

nate alone, as described in Materials and methods) when

calculating d34ST values in the products of thiosulfate

metabolism.

d34S of sulfate derived from thiosulfateoxidation by growing cultures ofH. neapolitanus strains C, X and FA11

Cultures of strain C (Table 1) and strain X in early

exponential growth (7.5 h after inoculation) had produced

sulfate equivalent to 11% and 9% of the initial thiosulfate-

sulfur. The sulfate was recovered and d34S found to be

15.8% and 14.2%, respectively. The 35S tracer tests showed

that this sulfate was derived almost exclusively from the

sulfonate-sulfur of thiosulfate, for which the d34S was

110.5%. Relative to the 34S/32S ratio of the sulfonate-sulfur

of thiosulfate (taking sulfonate-d34S = 0), the sulfate formed

from it thus had d34ST values of � 4.7% and � 6.2%,

respectively, for strains C and X. Significant fractionation in

favour of the lighter isotope thus occurred during sulfate

formation from the � SO3� group.

In another experiment in which all three strains were

sampled after 7–12 h growth, when 6.9–7.5% of the initial

thiosulfate had been converted to sulfate, the mean sulfate

d34S value for all three cultures was 14.1� 0.1%. The d34S

was recalculated on the basis that the sulfate was derived

wholly from the � SO3� group, and again indicated signifi-

cant fractionation in favour of 32S with a d34ST value of

� 6.3%.

34S/32S discrimination in the polythionates (andresidual thiosulfate) formed by growingcultures of H. neapolitanus strain C

In the experiment of Table 1, marked enrichment of 34S in

the sulfonate-residues of the combined thiosulfate 1 poly-

thionates was seen, with a clear fractionation of 32S into the

polythionate sulfane-sulfur after exhaustion of thiosulfate.

From 35S tracer experiments it was estimated that the

sulfane-sulfur of trithionate in the 13.5 h sample was derived

100% from the sulfane of thiosulfate, and that the trithio-

nate sulfonate-sulfur was about 60% from the sulfane of

thiosulfate. For tetrathionate, the distribution was about

2 : 1 in its sulfane and sulfonate groups from the sulfane of

thiosulfate. Sulfate recovered between 10.5 and 13.5 h of

culture had a d34S of 12.3% (Table 1), indicating selection

in favour of 32S, whereas the precursor sulfonate groups in

this time period were showing d34S values rising to about

126% (Table 1).

In a separate experiment, a culture was sampled after 10 h

growth (when 24.2% of the initial thiosulfate-sulfur con-

centration remained), and sulfate (46.3% total-sulfur),

thiosulfate, trithionate (19.3%) and tetrathionate (10.2%)

were recovered individually by ion exchange chromatogra-

phy, and 34S/34S ratios determined for each. d34ST values for

each entire molecule relative to the 34S/32S ratio of the whole

substrate-thiosulfate ion (as d34ST = 0), were 14.1%(thiosulfate), � 2.8% (sulfate), � 3.8% (trithionate) and

13.8% (tetrathionate). At this intermediate stage, thiosul-

fate and tetrathionate thus showed similarly raised d34ST

values, whereas trithionate and sulfate showed depletion in

their overall 34S content.

d34S of sulfate derived from thiosulfateoxidation by nongrowing suspensions ofH. neapolitanus strains C

The time course, extent and products of thiosulfate oxida-

tion, were assessed by simultaneous incubation of duplicate

flasks with unlabelled thiosulfate (for 34S/32S and direct

chemical analyses), or with [32S� 35SO3]2�or [35S� 32SO3]2�,

and in Warburg flasks (see Materials and methods). After

25 min incubation, oxygen uptake in the Warburg control

flasks was 23% of that required for complete oxidation of

the thiosulfate to sulfate; and by chemical analysis and 35S-

assessment, about 20% of the thiosulfate-sulfur was recov-

ered as sulfate. From the 35S analyses, 35% of the sulfate was

derived from the sulfane of thiosulfate and 65% from the

sulfonate group. The d34S of this sulfate was 14.3%,


302 D.P. Kelly

showing depletion in 34S relative to the d34S of the initial

thiosulfate ion (17.0%). Calculating d34ST for the sulfate

based on the 34S/32S ratio of a 65 : 35 origin from the

thiosulfate sulfonate- and sulfane-sulfurs, gave a d34ST

value of � 4.0%. Thus, nongrowing suspensions of

H. neapolitanus showed similar discrimination in favor of32S in the early stages of sulfate formation as did growing

cultures.

Discussion

The data demonstrate that significant fractionation in favor

of 32S occurs during the oxidation to sulfate of the sulfonate-

sulfur of thiosulfate by growing cultures of three strains of

H. neapolitanus, and that considerable enrichment of 34S

into the sulfonate-sulfur of polythionates occurs, especially

in the later stages of oxidation. The 35S-tracer control

experiments enabled the approximate origins of the sulfur

in the compounds assayed for their 34S/32S ratios to be

determined. These showed that there was depletion of 34S in

the sulfane-sulfur during thiosulfate oxidation and poly-

thionate (especially trithionate) formation, and progressive

enrichment of 32S in sulfate as sulfonate-sulfur was oxidized.

Two phenomena thus contributed to isotope discrimina-

tion: preferential conversion to sulfate of the 32S-sulfonate of

thiosulfate and accumulation later in oxidation of 34S in the

polythionate sulfonate groups, with preferential cleavage of

these to yield 32S-enriched sulfate. Metabolism of the

sulfane-sulfur of the original thiosulfate was thus a key

factor in isotope discrimination as oxidation proceeded,

and thiosulfate was resynthesized from its own sulfane-

sulfur. Chemically, there is known to be no isotopic ex-

change between the sulfane- and sulfonate-sulfur atoms of

thiosulfate, and no exchange between sulfite and the sulfo-

nate-sulfur of thiosulfate (Ames & Willard, 1951), hence

conversion of the sulfane of thiosulfate to sulfonate could

occur only by an enzyme-catalyzed mechanism (Kelly &

Syrett, 1966).

Sulfate enriched with 32S was formed in the later stages

of exponential growth (Table 1: 10.5–13.5 h), when thiosul-

fate became exhausted and about half of the total sulfate

production occurred (rising from 33% to 64% of the

initial thiosulfate-sulfur between 10.5–13.5 h; Table 1). The

sulfate present at 13.5 h had a d34S of 12.3 %. This can

be correlated with the increasing proportion of the

original sulfane-sulfur in thiosulfate appearing in the sulfo-

nate position. At 10.5 h, about 70% of the thiosulfate

molecule was sulfane-derived, and by 13.0–13.5 h all the

thiosulfate originated from the original sulfane-sulfur. The

sulfate formed between 10.5–13.5 h, arose from the original

sulfane- and sulfonate-sulfur in a ratio of about 4 : 1,

indicating a d34ST value for the sulfate of about � 3.3%,

which can be compared with the d34ST values of � 4.7%to � 6.3% seen in the early stage of growth, when most

sulfate was formed from the original sulfonate-sulfur of

thiosulfate. The proportional fractionation in favour of32S was thus fairly constant throughout the exponential

growth period. Similarly, the results showed that (after

exhaustion of thiosulfate by 13.5 h; Table 1) about 75% of

the polythionate-sulfur was derived from the sulfane-atom

Table 1. d34S determinations for sulfate and the sulfane- [� S� ] and sulfonate- [� SO3�] sulfur atoms of sulfur compounds during the growth of

Halothiobacillus neapolitanus strain C on thiosulfate

Time (h)

Sulfur compound concentrations (% of total sulfur recovered)

S2O32� SO4

2� S3O62� S4O6

2�

0 100 0 0 0

7.5 81 11 8 0

10.5 51 33 12 4

13.5 2 64 20 14

Time (h)

d34S (%) of sulfate and the [� S� ] and [� SO3�] groups of the combined thiosulfate, trithionate and tetrathionate recovered

Sulfate [Sulfane-sulfur] [Sulfonate-sulfur]�

7.5 15.8 [� 4.7]w 14.8 [10.4] 111.7 [11.2]

10.5 16.1 14.7 [10.3] 114.8 [13.5]

13.5 12.3 10.8 [� 3.6] 125.7 [119.9]

�d34S values for sulfonate-sulfur samples were calculated by difference between d34S values for sulfate alone and sulfate 1 total sulfonate recovered

after mercury degradation (see Methods and materials), relative to their respective concentrations.wNumbers in parentheses are estimated d34ST values relative to the measured or calculated d34S values of the precursor sulfur atoms for these groups at

the given sample times (see further details in the text).

A culture (1.2 L) was sampled at the times shown, centrifuged, and the supernates analyzed to recover sulfate and to degrade the combined thiosulfate

and polythionates into their constituent sulfane- and sulfonate-sulfur groups for 32S/34S determinations.



of thiosulfate. This 3 : 1 ratio from the two substrate-

thiosulfate atoms was used to calculate the d34ST of the

sulfonate-sulfur of the accumulated polythionate as

119.9%. Conversely, the 32S enrichment of the polythionate

sulfane-sulfur at this time gave a d34ST of � 3.6%.

The mechanism of thiosulfate oxidation by obligately

chemolithotrophic Beta and Gammaproteobacteria is in-

completely understood, and unlike the Sox pathway of

Alphaproteobacteria (Friedrich et al., 2001, 2008), requires

both periplasmic and cytoplasmic enzymes, and involves

polythionate intermediates (Kelly & Syrett, 1966; Lu & Kelly,

1988a, b; Kelly, 1989; Beller et al., 2006). Some genes

encoding the a-proteobacterial Sox pathway have been

found in chemolithotrophic Beta and Gammaproteobacteria

(Petri et al., 2001; Meyer et al., 2007), including soxXYZAB,

but not soxCD, in the Thiobacillus denitrificans genome,

although the soxB gene is absent from Halothiobacillus

hydrothermalis strain HY-66 and from some isolates of

thiosulfate-oxidizing Thiomicrospira species (Petri et al.,

2001; Beller et al., 2006; Meyer et al., 2007). To date, only

the soxB gene has been reported in the H. neapolitanus strain

used in the present study (Petri et al., 2001; Meyer et al.,

2007), and its role in thiosulfate oxidation remains unclear.

The insignificant fractionation of 34S during thiosulfate

oxidation by Paracoccus versutus compared with the present

results with H. neapolitanus suggests that the organisms

differ in the routes by which they form sulfate from

thiosulfate.

The present state of our knowledge of the sulfur-oxidation

pathways in Gammaproteobacteria thus makes it premature

to interpret the isotope fractionation data in terms of the

alphaproteobacterial Sox model, and it is more appropriate

to apply the simplest mechanistic interpretation of the

isotope fractionation data that is consistent with results

obtained from H. neapolitanus. In the Sox model, sulfate

would be released directly from enzyme-bound thiosulfate

by the action of the sulfate thiol esterase SoxB/Protein B

enzyme (Lu & Kelly, 1983; Wodara et al., 1994; Bamford

et al., 2002; Sauve et al., 2007), as sulfite is not a free

intermediate in the Sox pathway (Kelly, 1989; Friedrich

et al., 2008). It is, however, probable that sulfite is a free

intracellular intermediate in thiosulfate oxidation by Beta

and Gammaproteobacteria, produced by the enzymatic clea-

vage of thiosulfate to [S] and SO32�, catalyzed by a sulfur-

transferase such as rhodanese, with the subsequent

oxidation of sulfite to sulfate being catalyzed by one or more

of sulfite dehydrogenase, adenylylsulfate (APS) reductase,

and reverse dissimilatory sulfite reductase (Kelly, 1989, 2003;

Kelly & Wood, 1994a–c; Dahl & Truper, 1994; Taylor, 1994;

Truper, 1994; Beller et al., 2006; Kappler, 2008). The scission

of thiosulfate to [S� ] and [� SO3] moieties would be

expected to favor release of 32S-sulfite, leaving thiosulfate

with 34S-enriched sulfonate-sulfur (Cypionka et al., 1998),

without a large change in the d34S of the sulfane atom, as

was observed experimentally (Table 1). Oxidation of sulfite

to sulfate favours the light isotope, resulting in enrichment

of 34S in the residual sulfite. The 35S-experiments showed

that thiosulfate was resynthesized from the [S� ] atom of

thiosulfate, which would have required oxidation of some

[S] to SO32� and oxidative condensation of [S] and sulfite to

form a thiosulfate ion (Kelly & Syrett, 1966; Kelly & Wood,

1994a). As oxidation to sulfate of the initial and resynthe-

sized thiosulfate ions progressed, there would be little

change in the d34S of the sulfane-atom but progressive

increase in the sulfonate-34S as it is discriminated against in

the scission to release of sulfite and further enriched because

of the increasing d34S of the sulfite available for thiosulfate

resynthesis. Trithionate formation (accounting for up to

20% of the initial thiosulfate-sulfur; Table 1) by oxidative

condensation of a thiosulfate ion with sulfite resulted in

increased enrichment of its sulfonate-sulfur with 34S because

of the progressive increase in the d34S of the available

thiosulfate-sulfonate and sulfite. These interpretations are

consistent with the d34S data (Table 1). Further fractionation

could occur when trithionate is recycled (via trithionate

hydrolase) to thiosulfate, with the release of sulfate (Trudin-

ger, 1964b; Kelly, 1989).

Tetrathionate arises by oxidative condensation of two

thiosulfate ions catalyzed by the tetrathionate synthase

enzyme (Kelly & Wood, 1994c):

½S2SO3�2� þ ½S2SO3�2� ) ½O3S2S2SO3�2� þ 2e�

This condensation is ‘d34S-neutral’ as thiosulfate and

tetrathionate are in very rapid isotopic equilibrium (Fava &

Bresadola, 1955), negating any 34S discrimination in the

oxidative condensation, and consistent with the absence of34S/32S fractionation when H. neapolitanus converts thiosul-

fate predominantly to tetrathionate under some conditions

(data not shown). Thiosulfate and tetrathionate will thus

always be in equilibrium with respect to d34S, whereas both

are present in solution.

Trithionate and tetrathionate thus make distinct contri-

butions to the isotope discrimination observed because

earlier work showed that tetrathionate does not appear to

be a precursor of trithionate, and reaction of thiosulfate with

tetrathionate, measured with 35S, to produce trithionate is

also too slow to explain trithionate formation (Fava, 1953;

Kelly & Syrett, 1966). The most likely mechanism for isotope

discrimination during trithionate formation involves the

enzyme-catalyzed combining of thiosulfate and sulfite, as

the alternative, purely chemical isotope exchange between

thiosulfate and trithionate, is extremely slow (Fava & Pajaro,

1954; Kelly & Syrett, 1966).

The demonstration of discrimination among stable sulfur

isotopes by this Gammaproteobacterium is consistent with


304 D.P. Kelly

the earlier observations of Kaplan & Rittenberg (1964), and

supports the geobiological role of such bacteria as the agents

of the sulfur isotope anomalies seen in some altered sulfide/

sulfate mineral deposits (Nissenbaum & Rafter, 1967).

Acknowledgements

I am greatly indebted to the late Dr Athol Rafter (Rafter

Stable Isotope Laboratory, Institute for Nuclear Studies,

Lower Hutt, NZ) in whose laboratory all the stable isotope

measurements were made, and to Lyn Chambers who

carried out part of the experimental work at the Baas

Becking Geobiological Laboratory, Canberra, Australia. The

laboratory has now been disbanded, and Lyn Chambers

is retired. I am also very grateful to Prof. Colin Murrell and

Dr Ann Wood for several critical readings of the manuscript.

References

Ames DP & Willard JE (1951) The kinetics of the exchange of

sulfur between thiosulfate and sulfite. J Am Chem Soc 73:

164–172.

Bamford VA, Bruno S, Rasmussen T, Appia-Ayme C, Cheesman

MR, Berks BC & Hemmings AM (2002) Structural basis for

the oxidation of thiosulfate by a sulfur cycle enzyme. EMBO

J 21: 5599–5610.

Beller HR, Chain PSG, Letain TE, Chakicherla A, Larimer FW,

Richardson PM, Coleman MA, Wood AP & Kelly DP (2006)

The genome sequence of the obligately chemolithotrophic,

facultatively anaerobic bacterium Thiobacillus denitrificans.

J Bacteriol 188: 1473–1488.

Bottcher ME, Sievert SM & Kuever J (1999) Fractionation of

sulfur isotopes during dissimilatory reduction of sulfate by a

thermophilic gram-negative bacterium at 60 1C. Arch

Microbiol 172: 125–128.

Chambers LA & Trudinger PA (1979) Microbiological

fractionation of stable sulfur isotopes: a review and critique.

Geomicrobiol J 1: 249–293.

Chambers LA, Trudinger PA, Smith JW & Burns MS (1975)

Fractionation of sulfur isotopes by continuous cultures of

Desulfovibrio desulfuricans. Can J Microbiol 21: 67–72.

Cypionka H, Smock AM & Bottcher ME (1998) A combined

pathway of sulfur compound disproportionation in

Desulfovibrio desulfuricans. FEMS Microbiol Lett 166: 181–186.

Dahl C & Truper HG (1994) Enzymes of dissimilatory sulfide

oxidation in phototrophic sulfur bacteria. Meth Enzymol 243:

400–421.

Detmers J, Bruchert V, Habicht KS & Kuever J (2001) Diversity of

sulfur isotope fractionations by sulfate-reducing bacteria. Appl

Environ Microbiol 67: 888–894.

Fava A (1953) Reazioni di scambio isotopico tra politionati e

tiosolfato con S35. Gazz Chim Italiana 83: 87–97.

Fava A & Pajaro G (1954) Cinetique de l’echange isotopique

thiosulfate-trithionate. J Chim Phys 51: 594–599.

Fava A & Bresadola S (1955) Kinetics of the catalytic

rearrangement of tetrathionate. J Amer Chem Soc 77:

5792–5794.

Friedman I, Rafter TA & Smith GI (1995) A thermal, isotopic, and

chemical study of Lake Vanda and Don Juan Pond, Antarctica.

Antarctic Res Ser 67: 47–74.

Friedrich CG, Rother D, Bardichewsky F, Quentmeier A & Fischer

J (2001) Oxidation of reduced inorganic sulfur compounds by

bacteria: emergence of a common mechanism? Appl Environ

Microbiol 67: 2873–2882.

Friedrich CG, Quentmeier A, Bardichewsky F, Rother D, Orawski

G, Hellwig P & Fischer J (2008) Redox control of

chemotrophic sulfur oxidation of Paracoccus pantotrophus.

Microbial Sulfur Metabolism (Dahl C & Friedrich CG, eds),

pp. 139–150. Springer, Berlin.

Fry B, Cox J, Gest H & Hayes JM (1986) Discrimination between34S and 32S during bacterial metabolism of inorganic sulfur

compounds. J Bacteriol 165: 328–330.

Fry B, Gest H & Hayes JM (1988) 34S/32S fractionation in sulfur

cycles catalyzed by anaerobic bacteria. Appl Environ Microbiol

54: 250–256.

Hoek J & Canfield DE (2008) Controls on isotope fractionation

during dissimilatory sulfate reduction. Microbial Sulfur

Metabolism (Dahl C & Friedrich CG, eds), pp. 273–284.

Springer, Berlin.

Jensen ML (1965) Bacteriogenic sulfur isotope ratios in geology.

Beitr Mineral Petrogr 11: 405–414.

Kaplan IR & Rafter TA (1958) Fractionation of stable isotopes of

sulfur by thiobacilli. Science 127: 517–518.

Kaplan IR & Rittenberg SC (1964) Microbiological fractionation

of sulfur isotopes. J Gen Microbiol 34: 195–212.

Kappler U (2008) Bacterial sulfite-oxidizing enzymes – enzymes

for chemolithotrophs only? Microbial Sulfur Metabolism (Dahl

C & Friedrich CG, eds), pp. 151–169. Springer, Berlin.

Kappler U, Friedrich CG & Truper HG (2001) Evidence for two

pathways of thiosulfate oxidation in Starkeya novella (formerly

Thiobacillus novellus). Arch Microbiol 175: 102–111.

Karavaiko GI, Miller YM, Kapustin OA & Pivovarova TA (1981)

Fractionation of stable isotopes of sulfur during its oxidation

by Thiobacillus ferrooxidans. Microbiology (Moscow) 35:

667–671.

Kelly DP (1968) Fluoroacetate toxicity in Thiobacillus

neapolitanus and its relevance to the problem of obligate

chemoautotrophy. Arch Mikrobiol 61: 59–76.

Kelly DP (1989) Physiology and biochemistry of unicellular

sulfur bacteria. Autotrophic Bacteria (Schlegel HG & Bowien B,

eds), pp. 193–217. Science Tech Publisher, Madison, WI.

Kelly DP (1999) Thermodynamic aspects of energy conservation

by chemolithotrophic sulfur bacteria in relation to the sulfur

oxidation pathways. Arch Microbiol 171: 219–229.

Kelly DP (2003) Microbial inorganic sulfur oxidation: the APS

pathway. Biochemistry and Physiology of Anaerobic Bacteria

(Ljungdahl LG, Adams MW, Barton LL, Ferry JG & Johnson

MK, eds), pp. 205–219. Springer-Verlag, New York.



Kelly DP & Syrett PJ (1966) [35S]thiosulphate oxidation by

Thiobacillus strain C. Biochem J 98: 537–545.

Kelly DP & Wood AP (1994a) Whole organism methods for

inorganic sulfur oxidation by chemolithotrophs and

photolithotrophs. Meth Enzymol 243: 510–520.

Kelly DP & Wood AP (1994b) Synthesis and determination of

thiosulfate and polythionates. Meth Enzymol 243: 475–501.

Kelly DP & Wood AP (1994c) Enzymes involved in

microbiological oxidation of thiosulfate and polythionates.

Meth Enzymol 243: 501–510.

Kelly DP & Wood AP (1998) Microbes of the sulfur cycle.

Techniques in Microbial Ecology (Burlage RS, Atlas R, Stahl D,

Geesey G & Sayler G, eds), pp. 31–57. Oxford University Press,

New York.

Kelly DP & Wood AP (2005) Genus I. Halothiobacillus Kelly and

Wood. Bergey’s Manual of Systematic Bacteriology, Vol. 2.

(Brenner DJ, Krieg NR, Staley JT & Garrity GM, eds),

pp. 58–59, part B, Springer, New York.

Kelly DP, Shergill JK, Lu W-P & Wood AP (1997) Oxidative

metabolism of inorganic sulfur compounds by bacteria.

Antonie v Leeuwenhoek 71: 95–107.

Kelly DP, Wood AP & Stackebrandt E (2005) Genus II.

Thiobacillus Beijerinck. Bergey’s Manual of Systematic

Bacteriology, Vol. 2. (Brenner DJ, Krieg NR, Staley JT & Garrity

GM , eds), pp. 764–770, part C, Springer, New York.

Knoller K, Vogt C, Richnow H-H & Weise SM (2006) Sulfur and

oxygen isotope fractionation during benzene, toluene, ethyl

benzene, and xylene degradation by sulfate-reducing bacteria.

Environ Sci Technol 40: 3879–3885.

Krouse HR & Coplen TB (1997) Reporting of relative sulfur

isotope-ratio data. Pure Appl Chem 69: 293–295.

Lu W-P & Kelly DP (1983) Purification and some properties of

two principal enzymes of the thiosulphate-oxidizing

multienzyme system from Thiobacillus A2. J Gen Microbiol

129: 3549–3564.

Lu W-P, Swoboda BEP & Kelly DP (1985) Properties of the

thiosulphate-oxidizing multi-enzyme system from

Thiobacillus versutus. Biochim Biophys Acta 828: 116–122.

Lu W-P & Kelly DP (1988a) Kinetic and energetic aspects of

inorganic sulphur compound oxidation by Thiobacillus

tepidarius. J Gen Microbiol 134: 865–876.

Lu W-P & Kelly DP (1988b) Cellular location and partial

purification of the ‘‘thiosulphate-oxidizing enzyme’’ and

‘‘trithionate hydrolase’’ from Thiobacillus tepidarius. J Gen

Microbiol 134: 877–885.

Meyer B, Imhoff JF & Kuever J (2007) Molecular analysis of the

distribution and phylogeny of the soxB gene among sulfur-

oxidizing bacteria – evolution of the Sox sulfur oxidation

enzyme system. Environ Microbiol 9: 2957–2977.

Nissenbaum A & Rafter TA (1967) Sulfur isotopes in altered

pyrite concretions from Israel. Israel J Petrol 37: 961–962.

Petri R, Podgorsek L & Imhoff JF (2001) Phylogeny and

distribution of the soxB gene among thiosulfate-oxidizing

bacteria. FEMS Microbiol Lett 197: 171–178.

Quentmeier A, Hellwig P, Bardichewsky F, Grolle G, Kraft R &

Friedrich CG (2003) Sulfur oxidation in Paracoccus

pantotrophus: interaction of the sulfur-binding protein SoxYZ

with the dimanganese SoxB protein. Biochem Biophys Res

Commun 312: 1011–1018.

Quentmeier A, Janning P, Hellweg P & Friedrich CG (2007)

Activation of the heterodimeric central complex SoxYZ of

chemotrophic sulfur oxidation is linked to a conformational

change and SoxY-Y interprotein disulfide formation. Biochem

46: 10990–10998.

Reijerse EJ, Sommerhalter M, Hellweg P, Quentmeier A, Rother

D, Laurich C, Bothe E, Lubitz W & Friedrich CG (2007) The

unusual centers of SoxXA, a novel c-type heme-enzyme

essential for chemotrophic sulfur-oxidation of Paracoccus

pantotrophus. Biochemistry 46: 7804–7810.

Sauve V, Bruno S, Berks BC & Hemmings AM (2007) The SoxYZ

complex carries sulfur cycle intermediates on a peptide

swinging arm. J Biol Chem 282: 23194–23204.

Smock AM, Bottcher ME & Cypionka H (1998) Fractionation of

sulfur isotopes during thiosulfate reduction by Desulfovibrio

desulfuricans. Arch Microbiol 169: 460–463.

Taylor BF (1994) Adenylylsulfate reductases from thiobacilli.

Meth Enzymol 243: 393–400.

Trudinger PA (1964a) Oxidation of thiosulphate by intact cells of

Thiobacillus X: effects of some experimental conditions. Aust

J Biol Sci 17: 738–751.

Trudinger PA (1964b) The metabolism of trithionate by

Thiobacillus X. Aust J Biol Sci 17: 459–468.

Truper HG (1994) Reverse siroheme sulfite reductase from

Thiobacillus denitrificans. Meth Enzymol 243: 422–426.

Vairavamurthy A, Manowitz B, Luther G & Leon Y (1993)

Oxidation states of sulfur in thiosulfate and implications for

anaerobic energy metabolism. Geochim Cosmochim Acta 57:

1619–1623.

Van der Heijde HB & Aten AHW (1952) Tracer studies on some

reactions of thiosulphate and tetrathionate. J Am Chem Soc 74:

3706.

Wodara C, Kostka S, Egert M, Kelly DP & Friedrich CG (1994)

Identification and sequence analysis of the soxB gene essential

for sulfur oxidation of Paracoccus denitrificans BG17.

J Bacteriol 176: 6188–6191.

Wood AP & Kelly DP (1986) Chemolithotrophic metabolism of

the newly-isolated moderately thermophilic, obligately

autotrophic Thiobacillus tepidarius. Arch Microbiol 144: 71–77.


306 D.P. Kelly

stable sulfur isotope fractionation and discrimination between the sulfur atoms of thiosulfate...

Documents